Documentation for modified -new- QRF model

Background

This is the skeleton rmd outlining the development of a revised (“new”) QRF model (See et al. 2021). This will eventually highlight the covariate selection process, QRF model fit, the RF extrapolation model, and finally compare extrapolation estimates between the original and new models for eight watersheds within the Upper Salmon Basin.

Covariate selection

Explain covariate selection process. DASH/resto compatible, high MIC, low correlation, missing data, lots of zeros, species specific models, and eventually converting to joint models.

QRF Model fit

Talk about relative importance and pdp plots.

Extrapolation model

Talk about GAA’s

Extrapolation comparison

Extrapolations of habitat capacity for Chinook salmon, by life-stage, within the North Fork Salmon watershed.

Extrapolations of habitat capacity for steelhead, by life-stage, within the North Fork Salmon watershed..

QRF Habitat Capacity Estimates

Comparison with previous extrapolation

Chinook Salmon

Change in predicted habitat capacity for Chinook salmon compared to the previous extrapolation, by life-stage, within the North Fork Salmon watershed.

Estimated chinook capacities and comparison with previous random forest extrapolations for eight watersheds

Model	Watershed	Predicted capacity	Capacity % change	Predicted capacity SE	SE % change
Juv summer	EF Salmon	689,267.0	34	74,826	141
Juv summer	Lemhi	532,806.3	-7	23,315	33
Juv summer	NF Salmon	184,759.5	-5	16,204	52
Juv summer	Pahsimeroi	200,435.3	-11	8,313	-1
Juv summer	Panther Cr	505,906.7	4	41,601	77
Juv summer	Upper Salmon	973,874.1	37	114,743	120
Juv summer	Valley Cr	467,674.0	13	72,134	117
Juv summer	Yankee Fork	524,567.5	54	90,995	281
Juv winter	EF Salmon	34,428.5	-75	67,228	388
Juv winter	Lemhi	61,977.1	-60	28,701	325
Juv winter	NF Salmon	24,708.7	-55	25,271	401
Juv winter	Pahsimeroi	33,654.4	-64	15,138	123
Juv winter	Panther Cr	60,186.4	-61	50,357	402
Juv winter	Upper Salmon	38,599.2	-83	78,869	224
Juv winter	Valley Cr	23,612.5	-82	47,311	256
Juv winter	Yankee Fork	29,059.3	-71	65,686	408
Redds	EF Salmon	363.3	-22	29,624	101,387
Redds	Lemhi	323.3	-4	14,345	155,682
Redds	NF Salmon	151.1	-14	9,143	112,226
Redds	Pahsimeroi	125.6	12	6,458	169,943
Redds	Panther Cr	392.8	-15	19,546	101,478
Redds	Upper Salmon	517.6	-28	36,800	76,211
Redds	Valley Cr	347.2	-38	22,465	64,316
Redds	Yankee Fork	394.2	-44	26,929	47,844

Steelhead

Change in predicted habitat capacity for steelhead compared o the previous extrapolation, by life-stage, within the North Fork Salmon watershed..

Estimated winter steelhead capacities and comparison with previous random forest extrapolations for eight watersheds

Model	Watershed	Predicted capacity	Capacity % change	Predicted capacity SE	SE % change
Juv summer	EF Salmon	225,462.7	-40	30,958	248
Juv summer	Lemhi	294,099.9	-26	19,878	190
Juv summer	NF Salmon	202,904.2	-23	23,114	204
Juv summer	Pahsimeroi	148,087.9	-32	9,877	61
Juv summer	Panther Cr	230,228.7	-25	16,642	164
Juv summer	Upper Salmon	213,209.3	-44	33,036	199
Juv summer	Valley Cr	154,009.8	-45	23,262	160
Juv summer	Yankee Fork	174,368.3	-46	30,533	405
Juv winter	EF Salmon	180,845.7	-54	54,038	91
Juv winter	Lemhi	228,750.3	-42	35,943	99
Juv winter	NF Salmon	163,163.2	-48	38,960	45
Juv winter	Pahsimeroi	129,045.6	-40	16,943	47
Juv winter	Panther Cr	205,509.7	-35	25,593	61
Juv winter	Upper Salmon	155,401.5	-63	62,129	90
Juv winter	Valley Cr	130,227.9	-61	45,613	61
Juv winter	Yankee Fork	150,547.5	-64	59,730	44
Redds	EF Salmon	367.3	-23	31,276	113,515
Redds	Lemhi	382.3	-5	22,932	139,872
Redds	NF Salmon	287.0	-20	21,118	116,877
Redds	Pahsimeroi	152.4	-22	12,285	137,702
Redds	Panther Cr	279.7	-18	15,998	121,356
Redds	Upper Salmon	393.2	-23	33,945	100,246
Redds	Valley Cr	311.7	-32	23,919	86,603
Redds	Yankee Fork	393.8	-34	30,994	96,783

Capacity by stream

Chinook

Steelhead